Background and purpose A recent retrospective study using an online list service established by the American Academy of Neurology has suggested that ischaemic cerebrovascular events may occur in patients who undergo ‘bubble studies’ (BS) with either transcranial Doppler (TCD) or transoesophageal echocardiography (TOE). The safety of TCD-BS for right to left shunt (RLS) identification was evaluated prospectively in an international multicentre study.
Methods Consecutive patients with cerebral ischaemia (ischaemic stroke or transient ischaemic attack (TIA)) were screened for potential ischaemic cerebrovascular events following injection of microbubbles during TCD-BS for identification of RLS at three tertiary care stroke centres. TCD-BS was performed according to the standardised International Consensus Protocol. TOE-BS was performed in selected cases for confirmation of TCD-BS.
Results 508 patients hospitalised with acute cerebral ischaemia (mean age 46±12 years, 59% men; 63% ischaemic stroke, 37% TIA) were investigated with TCD-BS within 1 week of ictus. RLS was identified in 151 cases (30%). TOE-BS was performed in 101 out of 151 patients with RLS identified on TCD-BS (67%). It was positive in 99 patients (98%). The rate of ischaemic cerebrovascular complications during or after TCD-BS was 0% (95% CI by the adjusted Wald method: 0–0.6%). Structural cardiac abnormalities were identified in 38 patients, including atrial septal aneurysm (n=23), tetralogy of Fallot (n=1), intracardiac thrombus (n=2), ventricular septal defect (n=3) and atrial myxoma (n=1).
Conclusion TCD-BS is a safe screening test for identification of RLS, independent of the presence of cardiac structural abnormalities.
Statistics from Altmetric.com
Patients with ischaemic stroke and transient ischaemic attacks (TIAs) with paradoxical embolism as the potential underlying stroke mechanism can be investigated using the transcranial Doppler (TCD) ‘bubble test’ (BS) to detect the presence of a right to left (RLS) shunt.1 Although transoesophageal echocardiography (TOE) is considered the gold standard for RLS detection across a patent foramen ovale, contrast enhanced TCD is no less sensitive, much better tolerated and moreover it allows easier quantification of RLS in a recumbent or standing position.2 3
A recent retrospective study using an online list service established by the American Academy of Neurology has suggested that ischaemic cerebrovascular events may occur in patients who undergo BS with either TCD or TOE.4 The authors documented three ischaemic strokes (IS) and two TIAs as cerebrovascular complications of 3314 BS performed in four certified stroke centres across the USA. They emphasised the absence of a validated protocol for performance of BS and suggested that additional study is indicated to assess the true prevalence of cerebrovascular complications after BS. We evaluated prospectively the safety of TCD-BS for RLS identification in an international multicentre study.
We prospectively evaluated consecutive patients with IS or TIA due to suspected paradoxical embolism (sudden onset of neurological deficits with or without a history of Valsalva manoeuvre before symptom onset, negative carotid duplex evaluation for haemodynamically significant carotid or vertebral artery disease, normal electrocardiogram) referred to the neurovascular laboratories at three tertiary care stroke centres (Alexandroupolis, Athens and Singapore) with experience in TCD-BS.5 6 All patients were included in a computerised data bank and their demographic characteristics and vascular risk factors were documented, as previously described.5 6 TCD-BS was performed at all centres according to the standardised International Consensus Protocol (ICP).7 In brief, we used 2 MHz motion mode TCD (ST3, Spencer Technologies, Seattle, Washington, USA; SONARA, Viasys Healthcare, Conshohocken, Pennsylvania, USA) to detect the middle cerebral artery (MCA) at a depth range of 40–60 mm (figure 1). Patients with absent temporal window were excluded from the present study. Sonographers were certified by the American Society of Neuroimaging in TCD procedures/interpretation.5 6 At the beginning of contrast TCD testing, we used agitated saline (mixture of 9 ml saline solution and 1 ml air).7 Agitation was performed through a three way stopcock and was injected into the right antecubital vein. Valsalva manoeuvre was initiated 5 s after agitated saline injection for 5 s and monitoring was continued for 1 min. Valsalva manoeuvre was considered optimal when MCA velocities decreased by ≥25%.7 The RLS was quantified on TCD using the ICP criteria7 on the basis of the maximum number of microbubble signals in the MCA in any single spectral frame: negative (no microbubbles), grade I (1–20 microbubbles), grade II (>20 microbubbles or ‘shower’ appearance of microbubbles) and grade III (‘curtain’ appearance of microbubbles).
In all centres there was an independent investigator (attending neurologist) documenting potential adverse events (including IS and TIA) during and after (20–60 min following contrast injection) TCD-BS. Additionally, all patients were asked if they experienced a new episode of headache during or after TCD-BS. All recurrent strokes during hospitalisation and in particular the first 24 h following TCD-BS were prospectively documented.
TOE-BS was performed in selected cases for confirmation of TCD-BS in all three centres. All structural cardiac abnormalities observed on TOE were documented prospectively, as previously described.5 6 Cardiologists interpreted TOE studies without knowledge of TCD results, while TCD sonographers were blinded to the echocardiography results.
Continuous parametric and non-parametric data are presented as mean±SD or as median (IQR), respectively. Non-continuous variables are presented as percentages. The adjusted Wald method, which provides the best coverage for binomial CI with small samples, was used for computation of 95% CI. As in the previous report the authors documented only one adverse event following TCD-BS without the denominator (number of TCD-BS) being reported,4 we were not able to calculate the prevalence of expected adverse events for formal sample size calculation. We estimated that a sample of 500 patients would be needed to test the hypothesis that the rate of adverse cerebrovascular events following TCD-BS is 0.3% or greater with a power of 80.2% and a one tailed α value of 0.05 (only an effect in the expected direction (rate of adverse events exceeding the theoretical value of 0.3%) will be interpreted). Alternatively, a sample of 500 patients would enable us to test the hypothesis that the rate of adverse cerebrovascular events following TCD-BS is 0.5% or greater with a power of 85.6% and a two tailed α value of 0.05 (interpreting an effect in either direction). The Statistical Package for Social Science (SPSS Inc, V.13.0 for Windows) was used for statistical analyses.
A total of 508 patients hospitalised with acute cerebral ischaemia (mean age 46±12 years, 59% men; 63% stroke, 37% TIA) were investigated with TCD-BS within 1 week of ictus in the three collaborative centres. Demographic characteristics and stroke risk factors are shown in table 1. RLS was identified in 151 cases (30%) using TCD-BS (figure 1). TOE-BS was performed in 101 of 151 patients with RLS present on TCD-BS (67%). The presence of RLS was confirmed in 98% of the evaluated patients (n=99). In two cases the initial TOE-BS was negative for RLS. However, when the study was repeated (because of the positive TCD-BS), the presence of a ventricular septal defect (that had been overlooked at the original investigation) was visualised in one case. In the second patient the study became positive when the patient cooperated better during the Valsalva manoeuvre.
TCD and TOE findings are summarised in table 2. Structural cardiac abnormalities were identified in 38 patients, including atrial septal aneurysm (n=23), tetralogy of Fallot (n=1), intracardiac thrombus (n=2), ventricular septal defect (n=3) and atrial myxoma (n=1). The rate of ischaemic cerebrovascular complications during or after TCD-BS was 0% (95% CI by the adjusted Wald method: 0–0.6%). No case of headache was documented during or after the saline–air contrast injection. There was no case of recurrent stroke during the first 24 h following TCD-BS. No patient with paradoxical embolism due to pulmonary arteriovenous fistula was documented in our cohort.
Our multicentre study prospectively showed that TCD-BS is a safe screening test for identification of RLS, independent of the presence of cardiac structural abnormalities. More specifically, our study was powered to detect an incidence of adverse cerebrovascular events (following air–saline contrast injection) of 0.3% or greater and our findings (0 events) indicate with high confidence level (α value of 0.05) that TCD-BS has minimal or no complications. Our experience is very similar to that of the investigators who developed the ICP for TCD-BS.7 Interestingly, they stated that “at the dose recommended, there are currently no reports on side effects after air/saline administration”. To the best of our knowledge there is no other report in the literature of ischaemic cerebrovascular events being associated with TCD-BS, except for the recent retrospective study by Romero et al where one IS was documented following a TCD-BS.4 The investigators also recognised another four events after TOE/TTE-BS. Unfortunately, they provide only the denominator for the combined transthoracic echocardiographic, TOE and TCD studies (n=3313) and do not report specifically how many BS were performed separately using TCD or echocardiography. Our findings indicate with high confidence that the expected prevalence of adverse cerebrovascular events following air–saline contrast injection would be less than 0.3%.
Interestingly, it should be noted that previous reports have documented cerebrovascular complications during transoesophageal echocardiography BS.8 9 However, it can be argued that these ischaemic events may be causally related to the invasiveness of TOE (causing cardiac dysrhythmias, hypotension or, rarely, bacterial endocarditis),8 10 the presence of arterial catheters8 and not to the air–saline injection per se. The dose of 1 ml of air that is recommended in the ICP for TCD-BS may be considered safe because studies in animals have suggested that either a large bolus of air (20 ml) or small continuous amounts (11 ml/min) introduced into the venous system may generate intra-arterial bubbles able to cause embolism.11 Moreover, other investigators have shown in an animal experiment that 2 ml of air injected intra-arterially was the lowest dose needed to cause air embolism in a 7 kg macaque. By extrapolation, the critical volume for a 70 kg human should be well in excess of 2 ml, even for intra-arterial injections of air.12
Certain limitations of the present study need to be acknowledged. Firstly, TCD-BS negative patients were not evaluated with TOE and therefore the negative predictive value of TCD-BS for RLS detection could not be formerly evaluated. Secondly, our sample size was moderate (n=508) and further independent confirmation is needed by larger cohorts to increase the statistical confidence in our findings.
In conclusion, our findings underline the safety and importance in performing TCD-BS for RLS detection in a standardised manner using the currently recommended ICP. Additional prospective independent validation studies are needed to clarify the true incidence of paradoxical embolism following TCD-BS.
Competing interests None.
Ethics approval The ethics committees of all three institutions involved in the study approved the study protocol.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.